Apple Search Ads keyword and match type rules that hold up as you scale

An Apple Search Ads account gets harder to run at scale when the same search query is reachable from several places. Budget and bids end up competing with themselves, and nobody can say which keyword earned which install. The fix is less clever than it sounds: separate campaigns by intent, keep match types in their own lanes, and use negative keywords as the fence. Apple's own best-practices page describes this layout, and it is a good default.

Give each match type one job

Apple recommends four campaign types: brand, category, competitor and discovery. For the first three, Apple says to set keywords to exact match and make sure Search Match is off, with your most aggressive maximum cost-per-tap bids on your best performing keywords. Discovery is the only campaign that uses broad match and Search Match, in two ad groups: one with broad match keywords and Search Match off, and one with no keywords and Search Match on.

The logic is about control. Exact match tells you which term triggered the ad and lets you set a bid for it. Broad match and Search Match are for finding terms you have not thought of. Mixing them in one campaign means the loosest setting often spends the money, and you lose the ability to read the result.

Use negatives to keep discovery from stealing traffic

Apple says to add every keyword from your brand, category and competitor campaigns to discovery as exact match negatives. That lets discovery find related searches without bidding against the keywords you already manage. When discovery turns up a term that performs, the loop is simple:

1. Add the term to the right campaign as an exact match keyword.
2. Add the same term to discovery as an exact match negative.
3. Remove or lower the bid on any broad match keyword that was producing the same term.

Run this loop on a schedule, say every week or two, instead of whenever someone notices. An account maintained in bursts accumulates overlap.

Bid by the role of the term, not the average

Terms in different campaigns have different economics, so a single target does not fit. Brand terms usually convert best and cost least, because people already know the app. Category terms bring new users who are looking for a solution. Competitor terms carry the highest chance of mismatch, since the searcher wanted someone else. Set a target for each role and judge it separately. A competitor campaign judged against a brand campaign's cost per install will always look broken, and one judged only on its own cost per paying user may be worth keeping.

Hypothetically, suppose a brand campaign yields installs at $1, a category campaign at $3, and a competitor campaign at $5. If paid conversion is three times higher on category installs than on competitor installs, the category campaign is the better deal, even though both cost more than brand. That is why the decision metric should be cost per paying user, not cost per tap.

Scale by adding territory, not by loosening everything

When a campaign hits its budget, the instinct is to widen: more broad match, more Search Match, bigger bids. We would rather scale in this order:

1. Raise bids on exact match keywords that are budget-limited and profitable.
2. Add new exact match terms from discovery and from your own research on category language.
3. Add countries or regions. Apple suggests grouping markets with similar goals or shared traits, such as customer value or language, into one campaign, and creating separate campaigns for larger markets that need their own budget. Apple also notes that competitor keywords, Search Match and broad match may not suit every market, so test them market by market.
4. Increase discovery spend last, and only after the negatives are current.

Where to bend the rules

Small accounts with a handful of keywords may not need four campaigns. One brand campaign and one combined discovery campaign is a reasonable start, as long as brand terms are negated in discovery. Very large accounts may split category campaigns by intent, so that each group has its own bid and can be read separately. In either case, keep the rule that a given query has exactly one place it is meant to win.

Practical takeaway

Export your keyword and search term data, and for each query check whether it is reachable from more than one campaign. Where it is, assign one owner and negate it elsewhere. Do that before you raise any budget.

Sources

• Apple Search Ads, Campaign structure best practices